5 replies to this topic

[jcschultz01]

Strange behavor in Autostakkert! on long sequences of the lunar eclipse. 

See the attached screen capture of Autostakkert! on  the entire 7 minute segment and then reducing this to a 1 minute segment from the 7 minutes.

I had about 150 CR3 (had to convert to TIFF for Autostakkert!)  images taken from 1:02 to 1:09 AM during the lunar eclipse on 2025-03-14.

The Autostakkertt! Analyse seemed to work OK with at least 50% of the frames above 50% quality. 

 

 

But when I tried to stack over this long a duration, the image was blocky and chopped up.  Note I was not tracking the moon but even tracking the change in size would probably throw off Autostakkert!.

 

Anyway, thought others might experience the same issue

The solution was to just stack subsets, 1 minute duration worked for me though I did not go through an measure image quality vs. stacking frame duration. 

 

Attached Thumbnails

• 
• 

[happylimpet]

Planet mode? If so surface mode would probably work better.

 

Your statement about quality doesnt really mean anything. Half of all people are taller than average. A few years ago a UK politician said he wanted to make all schools above average.

Edited by happylimpet, 19 March 2025 - 02:50 PM.

[ToxMan]

How about providing a screenshot of the AutoStakkert settings?

[Borodog]

I am not surprised long sequences will not stack correctly. The image changes significantly over the course of 7 minutes.

[naztronomy]

Try using larger alignment points and select just a few points manually. 

 

I've had this issue when the APs are too many and too small. Fewer and larger may work for you. 

[LukaszLu]

It happened to me that even too strong a movement of the tube, probably caused by a strong gust of wind, forced me to apply a frame limit - so as to avoid stacking the shifted ones. Autoskakkert clearly cannot cope with changes, the nature or scale of which exceeds what its algorithm predicts.